In Defense of One-Vs-All Classification

By Ryan M. Rifkin, Aldebaro Barreto da Rocha Klautau Jr.. 2004

2,069 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Hyperband: A Novel Bandit-Based Approach to Hyperparameter Optimization (2016, methodology): “We use the multi-class regularized least squares classiﬁcation model which is known to have comparable performance to SVMs (Rifkin and Klautau, 2004; Agarwal et al., 2014) but can be trained signiﬁcantly faster.”
GP-Tree: A Gaussian Process Classifier for Few-Shot Incremental Learning (2021, methodology): “Some classic non-augmentation based methods include the Laplace approximation (Williams & Barber, 1998), expectation-propagation (Minka, 2001), and the least square approach (Rifkin & Klautau, 2004).”
Deep Neural Networks as Gaussian Processes (2017, methodology): “Formulating classiﬁcation as regression often leads to good results (Rifkin & Klautau, 2004).”
Multi-Label Prediction via Compressed Sensing (2009, background): “Pe rhaps the most straightforward approach is the well-known one-against-all reduction [3], bu t this can be too expensive when the number of possible labels is large (especially if applied to the power set of the label space [4]).”
Neural Networks and Deep Learning (2018, unknown): “It was shown in [401] that the one-against-all approach to generalizing multiple classes seems to be as effective as the tightly integrated multiclass variants.”
The Interplay of Optimization and Machine Learning Research (2006, methodology): “The authors develop a fast and frugal algorithm for learning rankings by comparing the predict graph with the feedback graph resulting in a QP with linear constraints.”
Least squares linear discriminant analysis (2007, methodology): “One of the directions for future work is to compare this new formulation with other related algorithms for multi-class classiflcation algorithms [10, 11, 27 ].”
What Does Classifying More Than 10, 000 Image Categories Tell Us? (2010, background): “This apparent breakdown for 1-vs-all with linear classifiers comes despite a consistent line of work promoting this general strategy for multi-class classification [22].”
Sparse coding-based correlaton model for land-use scene classification in high-resolution remote-sensing images (2016, unknown): “Both of these schemes are popular for image classification.(31,33,34,35) We compared them in the following experiment.”
Building an Intrusion Detection System Using a Filter-Based Feature Selection Algorithm (2016, methodology): “Rifkin and Klautau [42] demonstrated that the OVA technique was preferred over OVO.”
Learning with Augmented Class by Exploiting Unlabeled Data (2014, methodology): “OVR-SVM : One-vs-rest SVM is a powerful scheme for multi-class classiﬁcation (Rifkin and Klautau 2004).”
On the Decoding Process in Ternary Error-Correcting Output Codes (2010, background): “However, complex classifiers and optimization can make the results obtained by the two former strategies comparable [11].”
Learning to detect unseen object classes by between-class attribute transfer (2009, background): “Multi-class classifiers are also often constructed as layers of binary decisions, from which the final output is inferred , e.g. [7] [28].”
A statistical approach to rule learning (2006, background): “-rest learning problems (Cohen, 1995; Rifkin and Klautau, 2004).”
Dynamic Feature Selection for Online Action Recognition (2013, methodology): “daBoost is a binary classifier whereas action recognitio e are different strategies for converting binary classif ne-vs-all (OVA) is computationally simple and as ac [20].”
Estimating accurate multi-class probabilities with support vector machines (2005, result): “But, a more recent paper [20] disagrees with these results.”
Label Embedding Trees for Large Multi-Class Tasks (2010, methodology): “All these methods are reported to perform similarly in terms of accuracy [25, 23].”